Remote sensing was used to assist in the quick and uniform classification of land use within the Marsh and Long Island Creek watersheds. Landsat5 satellite imagery was obtained depicting six different years ranging from 2003 to 2011. ERDAS Imagine 2010 image analysis software was used to perform the classification process. The Imagine software allows the remote sensing technician to choose between 6 different wavelengths (e.g. red, green, blue, near infra-red, etc.) to enhance an objects reflected radiation thereby enhancing the ability to identify objects within the image.
For this exercise the class chose to use bands 6, 4, & 2 to enhance the visibility of objects such as vegetation, asphalt, concrete and water. The classification process examined the entire image area (the watershed's limits) and automatically analyzed on a pixel-by-pixel basis in respect to its neighboring pixels to determine an intensity value that fell within one of twelve categories. These categories included items such as water, roads, forests, commercial, residential and agricultural. These categories and others were provided by the Georgia Adopt-a-stream program in coordination with the Atlanta Regional Commission.
These categories were chosen and used based on a predetermined quotient value for impervious surfaces. These particular surfaces do not allow water to quickly (or at all) penetrate into the ground. This simple action is of great concern to stream health. The water caught on these surfaces are corralled into storm drains and then gravity fed directly and quickly into a nearby stream. The speed at which this occurs is the second concern being addressed in this analysis. By choosing the above mentioned categories, the remote sensing analysis will help identify the percentage of area that is impervious.
Once the classification process was complete, each Landsat5 image was then used as the based of performing the spatial analysis to determine the percentage of each category. After the percentage of area for each category was obtained, it was then multiplied by the impervious surface quotient to help determine the amount of area potentially causing storm event run-off. ESRI's ArcMap Desktop 10.1 was used to further classify similar categories together while producing a standardized map template.
The maps and results from the analysis can be viewed here. An outline of the above mentioned steps including the equation used to determine percentages can be viewed also.
Remote Sensing Methods
Remote sensing was used to assist in the quick and uniform classification of land use within the Marsh and Long Island Creek watersheds. Landsat5 satellite imagery was obtained depicting six different years ranging from 2003 to 2011. ERDAS Imagine 2010 image analysis software was used to perform the classification process. The Imagine software allows the remote sensing technician to choose between 6 different wavelengths (e.g. red, green, blue, near infra-red, etc.) to enhance an objects reflected radiation thereby enhancing the ability to identify objects within the image.
For this exercise the class chose to use bands 6, 4, & 2 to enhance the visibility of objects such as vegetation, asphalt, concrete and water. The classification process examined the entire image area (the watershed's limits) and automatically analyzed on a pixel-by-pixel basis in respect to its neighboring pixels to determine an intensity value that fell within one of twelve categories. These categories included items such as water, roads, forests, commercial, residential and agricultural. These categories and others were provided by the Georgia Adopt-a-stream program in coordination with the Atlanta Regional Commission.
These categories were chosen and used based on a predetermined quotient value for impervious surfaces. These particular surfaces do not allow water to quickly (or at all) penetrate into the ground. This simple action is of great concern to stream health. The water caught on these surfaces are corralled into storm drains and then gravity fed directly and quickly into a nearby stream. The speed at which this occurs is the second concern being addressed in this analysis. By choosing the above mentioned categories, the remote sensing analysis will help identify the percentage of area that is impervious.
Once the classification process was complete, each Landsat5 image was then used as the based of performing the spatial analysis to determine the percentage of each category. After the percentage of area for each category was obtained, it was then multiplied by the impervious surface quotient to help determine the amount of area potentially causing storm event run-off. ESRI's ArcMap Desktop 10.1 was used to further classify similar categories together while producing a standardized map template.
The maps and results from the analysis can be viewed here. An outline of the above mentioned steps including the equation used to determine percentages can be viewed also.